Copyright1999 by the Genetics Society of America
Perspectives
Anecdotal, Historical and Critical Commentaries on Genetics
Edited by James F. Crow and William F. Dove
The Ascendency of Developmental Genetics, or How the T Complex Educated
a Generation of Developmental Biologists
Virginia E. Papaioannou
Department of Genetics and Development, College of Physicians and Surgeons of Columbia University, New York, New York 10032
T
HIS past year marked the 60th anniversary of the way of fitting “new” discoveries into a historical frame-work that forced one to realize that the same biological publication inGeneticsof a remarkable scientificpaper from a remarkable woman, Salome Gluecksohn- issues have been around for a long time. The seduction of technology may have led us to believe that we were Waelsch (then Gluecksohn-Schoenheimer), first
describ-ing the embryonic development of the tailless phenotypes posing novel questions and discovering things that could lead to enlightenment, but that first question resulting from the interaction of T (Brachyury) with
t “alleles” (Gluecksohn-Schoenheimer1938). In the from the audience reminded us of our intellectual in-heritance and the myriad contributions that simpler historical context of the discovery and investigation of
the T complex, that article does not represent the earli- methods and inductive reasoning had made in leading us to our point of departure in a new study. And so, using est beginnings; T and its interaction with t alleles was
discovered more than 10 years earlier (Dobrovolskaia- Salome’s 1938 article as a point of departure rather than a beginning, I offer some impressions on how
Zavadskaia1927), and the transmission ratio distortion
characteristic of t alleles was already known (Chesley the specter of the T complex affected a generation of mammalian geneticists, developmental biologists, and andDunn1936). Nor does it represent the first
embry-onic investigation of these genes; the embryembry-onic lethal- developmental geneticists.
The appropriately named T complex spawned an en-ity of T/T embryos had already been described (
Ches-ley 1932). For me, its significance is more symbolic. tire field of study that for years was plagued by complex genetics, bizarre interactions, contradictions, and misin-The article reported early (and beautiful) work from
one of several prominent women developmental biolo- terpretations. For those of you who might have been frightened out of the field early on, as I was for many gists who were highly influential role models for aspiring
women scientists, especially developmental biologists, years, and never had the time or inclination to get back to it, you will be happy to hear that the complexities throughout the following decades, and it deals with
subjects (embryology and the T locus) dear to my heart. have now been satisfactorily resolved by combinations of genetic, molecular, and evolutionary approaches, and Perhaps most importantly, it investigates a
phenome-non, the relationship between T and the tail interaction the accepted view can be quite simply explained (Silver
1985, 1993). Apart from the fact that t-bearing chromo-factor (also known as tct or t c omplex t ail interaction
factor), of the t-bearing chromosomes. This remains an somes have some peculiar rearrangements that took a vast amount of mouse breeding and genetics to under-outstanding aspect of the T complex about which we
have almost no clue and that we may not be much closer stand, the main problem that misled researchers for many years was the persistently pursued misinterpreta-to understanding now than we were 60 years ago, a
potent reminder that the T complex still has hidden tion that T and the many small t’s discovered in wild mice were alleles of the same locus, despite ample evi-secrets.
Salome has always been one to view scientific progress dence to the contrary. With that idea out of the way, the various aspects of T-complex phenotypes—effects in a historical context and always has pertinent examples
at the ready. Anyone who ever gave a research talk with on tail length, embryogenesis, fertility, transmission ra-tio, and recombination—can be explained as effects of Salome sitting in the front row expected, and possibly
dreaded, the inevitable moment when she got to her many different linked genes, rather than pleiotropic effects of a single locus. The complex chromosome itself feet with the first question and said something like “That
is very interesting and reminds me of an experiment is what needs explanation, and this can be done in evolutionary terms.
done by ‘X’ many years ago. . . .” Humility is not a usual
characteristic of research scientists, but Salome had a To paraphrase the eloquent explanations of Lee
It would not matter how the sterile mice were elimi-nated—simply that elimination occurred at an early time; in fact there are many different lethal genes in different t haplotypes that have effects at various stages of embryogenesis.
Over more than a million years, additional TRD genes and additional inversions were recruited to the t hap-lotypes, effectively increasing the genetic expanse of the T-complex region. In terms of its contribution to the evolution of the T complex, only the tail interaction between different t haplotypes and T, a mutation on the non-t-bearing or wild-type chromosome, remains to be explained. However, the T complex is far from solved and we are now in the early stages of the difficult task of identifying specific genes responsible for the charac-teristic features of the T complex, assigning them to chromosomal positions and determining their molecu-lar function.
With hindsight, it is easy to see why the T complex remained mysterious for so long. It was first described in the language of classical genetics at a time when the field of mammalian developmental biology and espe-cially mammalian developmental genetics barely ex-isted. It seemed to violate the most basic rule of inheri-tance, Mendel’s first law, and consequently to wreak havoc with Hardy-Weinberg. The mechanisms by which the rules might be broken generated speculations that were hard to prove or disprove. At the time, the field of developmental biology was dominated by studies of lower organisms, and the role of genes in controlling developmental mechanisms was largely ignored. The
Salome Gluecksohn-Waelsch.
discovery of the profound effects of T and the t alleles on developmental processes provided an early indica-tion that genes might play a role in controlling develop-ver (who was nedevelop-ver frightened by the problem;Silver
1985, 1993), the origin of the present-day t-bearing chro- mental mechanisms in vertebrates, but ironically, misin-terpretation and false assumptions led developmental mosomes, or haplotypes, can be explained by
postulat-ing the followpostulat-ing steps in their evolution. First, several biology down the wrong path for many years. It was at first an uneasy alliance between developmental biology loci that cause transmission ratio distortion (TRD
genes), favoring their own transmission to offspring and genetics.
Salome joined L. C. Dunn’s lab in 1933 after training and thus conferring a selective advantage to the
chro-mosome, occurred by chance in a chromosomal region. with Spemann in early amphibian development and began a career studying the complexities of gene actions With TRD gene linkage, selection would favor anything
that kept the genes together on this “selfish chromo- and interactions and their role on developmental mech-anisms in mammals. From the beginning, she and Dunn some.” Second, chromosomal inversions occurred, which
tended to keep the TRD genes together by suppressing recognized the paradox that the complex phenotypic manifestations and the existence of complementation recombination. The selective advantage of such
inver-sions should have rapidly led to fixation of this chromo- between the T/t alleles indicated separate loci, while other features, like the lack of recombination and the some, but the absence of fixation in modern mice is
evidence for a counterbalancing negative selection, spe- interaction affecting tail length, pointed to a single (complex) locus. Both possibilities were presented in a cifically, a detrimental effect of TRD genes on male
fertility that renders t/t mice sterile. Third, selective 1943 paper, but it was further stated that genotypically related effects could have a degree of independence in pressure to counterbalance the male sterility effect of
t/t mice—which would produce mice that were essen- development and that a thorough study of development “might reveal common sources from which diverse ef-tially evolutionary baggage, using resources without
leav-ing offsprleav-ing—favored the accumulation of unrelated, fects arise” (Dunn and Gluecksohn-Schoenheimer
1943, p. 39). These further caveats clearly indicate, to my embryonic lethal genes that essentially got rid of this
effort to merge developmental and genetic thinking. As major push toward this type of analysis (Papaioannou
1998). Genes can now be manipulated in a variety of mouse genetics began to expand and the battery of
mouse mutations affecting development increased, the ways, including overexpression, misexpression, and cre-ation of specific mutcre-ations by targeted mutagenesis, so T complex maintained a prominent place in the
devel-opment of ideas about how genes control develdevel-opment. that the genetic analysis of developmental events can be approached at multiple levels in the complex organism An interesting, two-part article in 1964 by Dunn and
Dorothea Bennett hints at the schizophrenia induced with the ultimate goal of assigning a genetic basis to developmental mechanisms.
by the T complex. In part I, Dunn discusses the t alleles
in terms of a juxtaposition of separate genes affecting For many years after it was discovered, the T complex fostered the development of concepts of genes as inter-related developmental processes that have persisted as
a block of genetic material (Dunn 1964). In part II, related functional units, moving the emphasis away from the gene as an independently acting unit and inte-Bennett, who became an influential and dynamic force
in mammalian developmental genetics, also used the grating ideas expounded by Hans Gruneberg on pedi-grees of causes in development. On the basis of the term “alleles” but considers the t alleles to be members
of the same chromosomal locus, making an assumption, assumption that the t alleles were structurally similar and specified membrane components, Bennett and col-on the basis of this belief, that they are all related in
genetic structure and therefore related in the processes leagues proposed an elegant, unifying theory of the T locus as a master developmental locus, explaining the they control (Bennett1964). Meanwhile, however, the
genetic evidence that the t alleles were not alleles at all varying time of death caused by different t alleles as effects on cell-cell interactions at successive critical but were different loci in an abnormal chromosomal
region was accumulating (Lyon and Philips 1959; stages of development (Bennett1964, 1975). This idea of the T locus unraveled, however, with the discovery
Lyon1960). Mary Lyon, with her incredible insight, was
making waves in genetic thinking with her explanation of inversions and the consequent suppression of recom-bination. With the first fine structure mapping of the of the phenomenon of X-chromosome inactivation in
female mammals that later became known as the Lyon region (Artztet al. 1982), attention shifted to the
indi-vidual genes included within the 15 cM comprising the hypothesis (Lyon 1961; Russell1961), but her clear
genetic evidence on the nature of the T complex was complex. To circumvent the problems associated with recombination suppression, ethylnitrosourea mutagen-largely ignored at the time.
The 1960s and the following decade saw tremendous esis screens were used as a means of recovering new alleles of known loci in the region (Bode1984) and of changes in the way vertebrate, especially mammalian,
developmental biology was approached. The develop- recovering the recessive lethal mutations (Shedlovsky
et al. 1986). Gradually, the different early lethal
pheno-ment of new experipheno-mental embryological techniques,
such as chimeras (Tarkowski1961;Mintz1964;Gard- types were all attributed to separate loci with no obvious structural or mechanistic connection apart from their
ner1968) and new and improved culture techniques
(Biggerset al. 1965;Brinster1970), made the mam- linkage and their embryonic lethality. There are at least 16 different t lethal effects (Klein et al. 1984), but,
malian embryo more accessible than ever before. There
was a strong impetus to apply the techniques and princi- as yet, no lethal allele has been cloned. The detailed phenotypic descriptions associated with many of these ples of molecular biology and to incorporate the insights
of microbial genetics as a means by which to understand lethal mutations are a treasure chest of information awaiting the identification of genes.
the developmental biology of more complex organisms
with vastly more complex genomes. The accumulating On the other hand, a candidate gene for the tail interaction effect, called T2, has been identified and wealth of information from experimental embryology
along with the accumulating wealth of naturally oc- cloned in the laboratory of Karen Artzt (Rennebeck
et al. 1998), a student of Dunn and Bennett and a
long-curring or induced developmental mutations in
mam-mals, especially the mouse, opened many pathways to time, influential expert in the T-complex field. This gene, identified by a transgene insertion, maps very a more direct inquiry into the roles of genes in directing
interrelated cascades of developmental events. Molecu- close to T, has similar effects on axial mesoderm forma-tion, but does not complement T. As there is no evi-lar genetic technology allowed the more direct
examina-tion of the link between gene expression and develop- dence for a direct effect of the transgene insertion on
T or its regulatory regions, this second gene could be
mental mechanisms. Now, with the increased power of
forward genetics and positional cloning, the classic tech- synonymous with the tail interaction factor, tct. Further analysis of this gene could soon get us closer to under-nique of analyzing a mutation by working from its
phe-notype to the gene to uncover the primary effects on standing this mysterious aspect of the T-complex and whether these separate genes act through similar mech-development could be supplemented by working in the
opposite direction, from the gene to the phenotype anisms.
The key to T itself came with the cloning of the gene
(see Davis and Justice 1998). The establishment of
(Herrmannet al. 1990; Herrmann1995) that is part to be key players in understanding the complexity of the developing embryo.
of a multigene family, the T-box family (Bollaget al.
In a Genetics Perspectives article written nearly a 1994). From this discovery, a new area of developmental
decade ago, Salome Gluecksohn-Waelsch discussed genetics, that has come to dominate my own research
Guido Pontecorvo’s and Barbara McClintock’s ideas on in the past few years, opened up. Previously, my meager
biological complexity and their application to the contribution to the study of the T complex consisted
T complex. She pointed out that “eventually clarifying of a single forgotten paper (Papaioannouet al. 1979)
the complexities that characterize the correlation be-resulting from a few weeks spent in Dorothea Bennett’s
tween the molecular identification of gene sequences lab teaching what I had learned from Wes Whitten about
in this interesting chromosomal region and their far-making aggregation chimeras and trying rather
unsuc-removed phenotypic expression” will require a great cessfully to understand t haplotypes. Only years later did
deal of patience and an analytical approach that takes I venture back to the field, albeit in an area peripheral to
into account the complexity of organisms (
Glueck-the T complex itself, spurred by Glueck-the possibility that Glueck-the
sohn-Waelsch1989, p. 724). This is no less true today, important developmental effects of T might herald a
when one considers that the cloning of the T gene led common feature of a conserved gene family with
multi-to the discovery of an entire gene family with many ple family members in the mouse.
developmental genes, adding a new level of complexity Uncovered on the basis of homology to the
DNA-to the attempt DNA-to elucidate the T complex. binding domain of the T-locus gene product, the T-box
I thank Lee Silver and Deborah Chapman for reading the
manu-gene family is a novel family of transcription factors.
script. This work was supported in part by the Raymond and Beverly
T-box genes have been identified in the genomes of a
Sackler Foundation and National Institutes of Health grant HD 33082.
wide range of metazoans from Caenorhabditis elegans to The photograph of Salome Gluecksohn-Waelsch was kindly provided human. At last tally, 10 different genes, in addition to by Vivian Gradus.
T, have been identified in the mouse and are found
scattered through the genome. From all indications, the
T-box gene family is implicated as playing critical roles LITERATURE CITED in development: first, by the well-studied mutant
pheno-Artzt, K., P. McCormickandD. Bennett,1982 Gene mapping
type of the proband of the family, T, then by the isolation within the T/t complex of the mouse. I. t-lethal genes are nonalle-of novel T-box genes from embryonic cDNA libraries lic. Cell 28: 463–470.
Bamshad, M., R. C. Lin, D. J. Law, W. S. Watkins, P. A. Krakowiak
and the demonstration of extensive embryonic
expres-et al., 1997 Mutations in human TBX3 alter limb, apocrine and
sion, and finally, in addition to the T mutations known genital development in ulnar-mammary syndrome. Nat. Genet.
16:311–315.
in many species, by the developmental phenotypes that
Basson, C. T., D. R. Bachinsky, R. C. Lin, T. Levi, J. A. Elkinset
result from other T-box gene mutations. Naturally
oc-al., 1997 Mutations in human TBX5 cause limb and cardiac
curring mutations in two T-box genes in humans are malformations in Holt-Oram syndrome. Nat. Genet. 15: 30–35.
Bennett, D.,1964 Abnormal association with a chromosomal
re-responsible for dysmorphic syndromes affecting the
de-gion in the mouse. II. Embryological effects of lethal alleles in
velopment of multiple organs (Bamshad et al. 1997;
the t-region. Science 144: 263–267.
Basson et al. 1997;Li et al. 1997). A zebrafish mutant Bennett, D.,1975 The T-locus of the mouse. Cell 6: 441–454.
Biggers, J. D., B. D. MooreandD. G. Whittingham,1965
Develop-found during a mutagenesis screen has a mutation in
ment of mouse embryos in vivo after cultivation from two-cell
a T-box gene that has drastic effects on the development ova to blastocysts in vitro. Nature 206: 734–735.
of somites (Griffinet al. 1998;Ruvinskyet al. 1998). A Bode, V. C.,1984 Ethylnitrosourea mutagenesis and the isolation of mutant alleles for specific genes located in the t region of the
targeted mutation of Tbx6 in mouse results in a dramatic
mouse chromosome 17. Genetics 108: 457–470.
developmental switch from somite to neural develop- Bollag, R. J., Z. Siegfried, J. A. Cebra-Thomas, N. Garvey, ment (ChapmanandPapaioannou1998). We are now E. M. Davisonet al., 1994 An ancient family of embryonically expressed mouse genes sharing a conserved protein motif with
able to apply all the power of reverse genetics to
identi-the T locus. Nat. Genet. 7: 383–389.
fying and investigating the developmental roles of all the Brinster, R. L.,1970 In vitro cultivation of mammalian ova. Adv. genes of this family. Working from gene to phenotype, Biosci. 4: 199.
Chapman, D. L.,andV. E. Papaioannou,1998 Three neural tubes
using targeted mutagenesis and other powerful
molecu-in mouse embryos with mutations molecu-in the T-box gene, Tbx6. Nature
lar techniques combined with the classic observational 391:695–697.
Chesley, P.,1932 Lethal action in the short-tailed mutation in the
skills of embryology, developmental geneticists will be
house mouse. Proc. Soc. Exp. Biol. Med. 29: 437–438.
able to define the critical developmental processes
af-Chesley, P.,andL. C. Dunn,1936 The inheritance of taillessness
fected by this gene family. One of the common features (anury) in the house mouse. Genetics 21: 525–536.
Crow, J. F.,1988 The ultraselfish gene. Genetics 118: 389–391.
of the family, DNA binding and presumably
transcrip-Davis, A. P., andM. J. Justice, 1998 An Oak Ridge legacy: the
tional regulatory activity, provides the basis for under- specific locus test and its role in mouse mutagenesis. Genetics standing common mechanisms in what are likely to be 148:7–12.
Dobrovolskaia-Zavadskaia, N.,1927 Sur la mortification
spon-myriad effects on development. Although the T-box
tane´e de la queue che la souris nouveau-ne´e et sur l’existence
genes may not be the master controllers of development d’un caracte`re (facteur) he´re´ditaire “non viable.” C. R. Seances
Soc. Biol. Fil. 97: 114–116.
Dunn, L. C.,1964 Abnormal association with a chromosomal region Lyon, M. F.,1961 Gene action in the X-chromosome of the mouse (Mus musculus L.). Nature 190: 372–373.
in the mouse. I. Transmission and population genetics of the
Lyon, M. F.,1991 The genetic basis of transmission-ratio distortion
t region. Science 144: 260–263.
and male sterility due to the t complex. Am. Nat. 137: 349–358.
Dunn, L. C.,andS. Gluecksohn-Schoenheimer (Waelsch),1943
Lyon, M. F.,andR. J. S. Philips,1959 Crossing-over in mice
hetero-Tests for recombination amongst three lethal mutations in the
zygous for t-alleles. Heredity 13: 23–32. house mouse. Genetics 28: 29–40.
Mintz, B.,1964 Formation of genetically mosaic mouse embryos,
Gardner, R. L.,1968 Mouse chimaeras obtained by the injection
and early development of ‘lethal (tt12/t12)-normal’ mosaics. J. Exp.
of cells into the blastocyst. Nature 220: 596–597.
Zool. 157: 273–292.
Gluecksohn-Schoenheimer, S., 1938 The development of two
Papaioannou, V. E.,1998 The coming of age of the transgenic era.
tailless mutants in the house mouse. Genetics 23: 573–584.
Int. J. Dev. Biol. 42: 841–846.
Gluecksohn Waelsch, S.,1989 In praise of complexity. Genetics
Papaioannou, V. E., K. Artzt, G. B. DooherandD. Bennett,1979
122:721–725.
Effects of chimaerism for two T/t complex mutations on gene
Griffin, K. J. P., S. L. Amacher, C. B. KimmelandD. Kimelman,
transmission ratio. Gamete Res. 2: 147–151. 1998 Molecular identification of spadetail: regulation of
zebra-Rennebeck, G., E. Lader, A. Fujimoto, E. P. LeiandK. Artzt,1998
fish trunk and tail mesoderm formation by T-box genes.
Develop-Mouse Brachyury the second (T2) is a gene next to classical T and ment 125: 3379–3388.
a candidate gene for tct. Genetics 150: 1125–1131.
Hermann, B. G.,1995 The mouse Brachyury (T) gene. Semin. Dev. Russell, L. B.,1961 Genetics of mammalian sex chromosomes.
Biol. 6: 385–394. Science 133: 1795–1803.
Hermann, B. G., S. Labiet, A. Poustka, T. KingandH. Lehrach, Ruvinsky, I., L. SilverandR. Ho,1998 Characterization of the
1990 Cloning of the T gene required in mesoderm formation zebrafish tbx16 gene and evolution of the vertebrate T-box family. in the mouse. Nature 343: 617–622. Dev. Genes Evol. 208: 94–99.
Klein, J., P. SiposandF. Figueroa,1984 Polymorphisms of t-com- Shedlovsky, A., J.-L. Gue´net, L. L. JohnsonandW. F. Dove,1986
plex genes in European wild mice. Genet. Res. 44: 30–46. Induction of recessive lethal mutations in the T/t-H2 region of
Li, Q. Y., R. A. Newbury-Ecob, J. A. Terrett, D. I. Wilson, A. R. J. the mouse genome by a point mutagen. Genet. Res. 47: 135–142.
Curtiset al., 1997 Holt-Oram syndrome is caused by mutations Silver, L. M.,1985 Mouse t haplotypes. Annu. Rev. Genet. 19: 179–208.
in TBX5, a member of the Brachyury (T) gene family. Nat. Genet. Silver, L. M.,1993 The peculiar journey of a selfish chromosome:
15:21–29. mouse t haplotypes and meiotic drive. Trends Genet. 9: 250–254.
Lyon, M. F.,1960 Effect of X-rays on the mutation of t-alleles in Tarkowski, A. K., 1961 Mouse chimaeras developed from fused
